DevFW-CICD/ingress-nginx-helm
14
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Merge pull request #4139 from choffmeister/fix/collect-metrics-if-metrics-per-host-false

Browse source 
Always collect metrics when --metrics-per-host=false
This commit is contained in:
Kubernetes Prow Robot 2020-08-08 12:02:19 -07:00 committed by GitHub
commit c500bd4b3f
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 8 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
docs/user-guide/monitoring.md
View file

@ -77,3 +77,9 @@ The username and password is `admin`
After the login you can import the Grafana dashboard from [official dashboards](https://github.com/kubernetes/ingress-nginx/tree/master/deploy/grafana/dashboards)
![Dashboard](../images/grafana.png)
## Caveats
### Wildcard ingresses
By default request metrics are labeled with the hostname. When you have a wildcard domain ingress, then there will be no metrics for that ingress (to prevent the metrics from exploding in cardinality). To get metrics in this case you need to run the ingress controller with `--metrics-per-host=false` (you will lose labeling by hostname, but still have labeling by ingress).

4
internal/ingress/metric/collectors/socket.go
View file

@ -230,8 +230,8 @@ func (sc *SocketCollector) handleMessage(msg []byte) {
}
for _, stats := range statsBatch {
if !sc.hosts.Has(stats.Host) {
klog.V(3).Infof("skiping metric for host %v that is not being served", stats.Host)
if sc.metricsPerHost && !sc.hosts.Has(stats.Host) {
klog.V(3).Infof("Skipping metric for host %v that is not being served", stats.Host)
continue
}